Article pubs.acs.org/ac
Novel Gas Chromatographic Detector Utilizing the Localized Surface Plasmon Resonance of a Gold Nanoparticle Monolayer inside a Glass Capillary Fong-Yi Chen, Wei-Cheng Chang, Rih-Sheng Jian, and Chia-Jung Lu* Department of Chemistry, National Taiwan Normal University, Taipei, 11677, Taiwan S Supporting Information *
ABSTRACT: This paper presents the design, assembly, and evaluation of a novel gas chromatographic detector intended to measure the absorbance of the localized surface plasmon resonance (LSPR) of a gold nanoparticle monolayer in response to eluted samples from a capillary column. Gold nanoparticles were chemically immobilized on the inner wall of a glass capillary (i.d. 0.8 mm, length = 5−15 cm). The eluted samples flowed through the glass capillary and were adsorbed onto a gold nanoparticle surface, which resulted in changes in the LSPR absorbance. The LSPR probing light source used a green light-emitting diode (LED; λcenter = 520 nm), and the light traveled through the glass wall of the capillary with multiple total reflections. The changes in the light intensity were measured by a photodiode at the rear of the glass capillary. The sensitivity of this detector can be improved by using a longer spiral glass capillary. The detector is more sensitive when operated at a lower temperature and at a slower carrier velocity. The calibration lines of 8 preliminary test compounds were all linear (R2 > 0.99). The detection limits (3σ) ranged from 22 ng (n-butanol) to 174 ng (2-pentanone) depending on the volatility of the chemicals and the affinity to the citrate lignads attached to the gold nanoparticle surface. This detector consumed a very low amount of energy and could be operated with an air carrier gas, which makes this detector a promising option for portable GC or μGC.
I
distance. The relationships between LSPR sensitivity, the distance from the surface, and the refractive index changes can be described using the following equation:20
n the past few decades, the localized surface plasmon resonance (LSPR) of metallic nanoparticles has been extensively studied for possible sensor applications by using its sensitivity toward environmental refractive index changes.1 The LSPR frequencies of many noble metal nanoparticles are within visible to near-IR range, which makes measuring them with a spectrometer simple. One of the most important applications of LSPR sensors is the label-free detection of biological species, which is a great simplification compared with traditional bioassay approaches.2−6 Many researchers are testing the parameters that affect the refractive index sensitivity of LSPR such as particle shape, size, and material.7−13 One particularly important factor is the sensitivity decay with distance from the surface of nanoparticles. Beeram and Zamborini reported a series of studies that used triangular nano-Au to investigate the distancedependent sensitivity and selective binding of biomolecules.14−16 Van Duyne and co-workers have estimated the sensitivity decay on silver nanotriangles using self-assembled monolayer in different chain lengths, and the decay distance for LSPR sensing was determined to be approximately 50 nm.17−19 Kedem et al. used layer-by-layer polymer deposition to determine that the sensing distance of spherical gold nanoparticles ranges between 10 and 20 nm and found the distance to be highly dependent on the nanoparticle size and shape.20 These results have profound implications on developing a LSPR biomolecule sensor because a portion of the large molecules (e.g., proteins, DNA) could be at a low-sensitivity © 2014 American Chemical Society
R = mΔη[1 − exp( −d /l)]
(1)
where R is the sensor response (i.e., wavelength shift or intensity), m is the sensitivity (i.e., Δλ or ΔI per refractive index unit, RIU), Δη is the change in the refractive index (RI), d is the thickness of the dielectric layer, and l is the plasmon decay length. When there is a homogeneous (d = ∞) change in the refractive index surrounding the nanoparticles, such as when they are immersed in a liquid with a given refractive index value, eq 1 can be simplified as R = mΔη. Gas-phase detections using LSPR are inherently difficult for two reasons: first, the adsorption on the surface of nanoparticles is weak for light gases or volatile organic vapors; second, the change in induced refractive indices per adsorbed molecule is small compared to the larger biomolecules (such as proteins) due to the effective volume occupied on the nanoparticle surroundings. In order to facilitate gas/vapor absorption on nanoparticles, Rubinstein and co-workers coated nanoparticles with a polymer film to enhance the vapor detection using LSPR.21 Chen et al. have reported the detection Received: October 1, 2013 Accepted: May 4, 2014 Published: May 4, 2014 5257
dx.doi.org/10.1021/ac4031829 | Anal. Chem. 2014, 86, 5257−5264
Analytical Chemistry
Article
of terpene vapors using thiolate-modified Au nanoparticles.22 Van Duyne’s group demonstrated the use of a high-resolution spectrometer to achieve the detection of gases (i.e., Ar, N2 or He) using the Ag nanotriangles.23 The same group later developed the coating of metal organic framework (MOF) molecules on Ag nanoparticles to enhance the sensitivity to light gases (e.g., CO2, CH4) and found a 14-fold enhancement in the response.24 Using the LSPR of metal oxide-coated nanoparticles for gas sensing has also been reported by Carpenter et al.25,26 In addition, a nanoantenna with a precisely controlled shape and distance has demonstrated superior sensitivity for H2 detection.27 Ibañez and co-workers reported the use of surfactant-stabilized gold nanoparticles as a highly durable LSPR sensor for VOCs.28 Very recently, vapor sensing using interparticle distance modulation in a 3-D plasmonic nanoparticle structure has been demonstrated by Potyrailo et al.29 Fan and Zellers used dual laser reflectance to probe the vapor-induced swelling of a nanoparticle film30 that had evolved from Fan’s previous design of a reflectance-type in-column GC detector.31 Our previous work has focused on studying various types of nanoparticles (e.g., Ag or Au nanospheres or shells) and surface modifications to investigate the selectivity of using LSPR as a VOC sensor array.32−34 In addition, Au nanoparticles have been used as a chemiresistor type of GC detector35,36 and in the stationary phase of a capillary column.37 In this work, we propose a new GC detector design that utilizes the LSPR of Au nanoparticles as a transducer mechanism. Instead of using a regular optical fiber, a reversed configuration (i.e., similar to hollow-fiber optics38,39) was used where gas flow and immobilized Au nanoparticles are inside a glass capillary and the incident light travels along the glass wall of the capillary. In order to minimize the size and improve the portability of the detector, a green light-emitting diode (LED) was used as the light source and a matched range photodiode was used as the light intensity detector. The LSPR detector was evaluated on a benchtop GC under various GC analytical conditions.
GNPs also influenced the gas sensing selectivity in later experiments. Detector Assembly. Both straight and spiral glass capillaries were tested in this study. The spiral glass capillaries were heated and bent by a local glass shop using the same straight glass capillaries obtained from Fishers Scientific. In some cases, the glass capillaries were coated with a silver mirror finish on the outside to maintain light intensity. The coating was performed by sealing the glass capillaries and immersing them in Tollens’ reagent. Amine-stabilized silver ions were gradually reduced by glucose in a basic solution and were used to coat the outer wall of the glass capillary. The photos of both silver-mirror coated and uncoated capillaries are shown in the Supporting Information (Figure S-1). The glass capillary was tightened between two 1/8 in. stainless Swagelok unions using stainless steel nuts and Teflon ferrules (Figure 1a). A short section of deactivated fused silica tube (i.d. 250 μm) was wrapped with a small piece of Teflon tape at one end and inserted into the glass capillary as the sample inlet for the LSPR detector. The deactivated silica tube was bent slightly allowing a green LED to align straight to the cross section of the glass capillary. A plastic syringe was used to
■
EXPERIMENTAL SECTION Nanoparticle Synthesis and Surface Modifications. A standard citric reduction approach has been used to synthesize gold nanoparticles (GNPs) with an average size of approximately 20 nm.32 The inner walls of the glass capillaries (od. = 2.0 mm, id. = 0.8 mm, Fishers Scientific, PA, USA) were first cleaned with Piranha solution and then rinsed with deionized water and dried. (CAUTION: “Piranha” solution reacts violently with organic materials; it must be handled with extreme care.) The glass capillaries were then filled with a solution of 10% (v/v) 3-aminopropyl trimethoxysilane (APTMS) in methanol and sealed for 1 h. After the inner surface of the capillary had been modified with APTMS, the capillary was rinsed with deionized water and filled with an aqueous solution of GNPs. The capillary was stored in a refrigerator for 2 h to allow the GNPs to bind to the glass surface. The preparation and characterization of this surface binding process was previously described by Natan’s group.40 The glass capillary was then rinsed with deionized water and ethanol repeatedly to remove unbounded GNPs and was kept dry for the detector assembly. The citrate that strongly adsorbed on the surface of GNPs as protector during aqueous phase synthesis was not removed completely by this mild rinsing procedure. Therefore, the citrate on the surface of the
Figure 1. (a) Scheme of LSPR-GC detector assembly. Photos of (b) 10 cm straight-type and (c) 15 cm spiral-type detectors. (Note: both pictures show the Ag layer coated on the outside of a glass capillary for better visualization.) 5258
dx.doi.org/10.1021/ac4031829 | Anal. Chem. 2014, 86, 5257−5264
Analytical Chemistry
Article
fix the positions of the LED light and the deactivated fused silica tube. Black vinyl electrical tape was wrapped around the outside of the plastic syringe to block the stray light from the environment. The rear end (i.e., sample exit) of the capillary was kept open and was aligned directly to a photodiode with a built-in filter for the range of the green light (S6429-01, Hamamatsu, Japan). The light was blocked by the fused silica tube at the center of the glass capillary. At the rear of the detector, the photodiode sensed the intensity of a green “halo” of light conducted through the glass wall of the capillary. Figure 1b,c shows the photos of both straight and spiral types of LSPR detectors after assembly had been completed. Instruments and Apparatus. A field emission scanning electron microscope (FESEM, JSM-6500F, JEOL) was used to examine the image of a GNP monolayer on a glass substrate. The SEM image is shown in the Supporting Information (Figure S-2). UV−vis spectra were measured using an Ocean Optics USB-2000 spectrometer. The initial determination of the LSPR spectrum for GNPs and the LSPR spectrum for a vapor response was performed with the same test configuration that was used in our previous report, for which GNPs were bonded to watch-glass slides in the cube cell of a UV−vis spectrometer.32 Test System for the LSPR GC Detector. The LSPR detector was tested on a benchtop gas chromatograph (HP5890, Agilent) equipped with a built-in flame ionization detector. During the comparison experiments with a thermal conductivity detector (TCD), another GC-TCD (Clarus 480, PerkinElmer, MA, USA) was used. A general-purpose DB-5 column (27 m, i.d. 0.53 mm, stationary phase thickness 1 μm) was used throughout all experiments. Test samples were prepared by injecting an aliquot of organic liquid into a Tedlar bag filled with a known volume of clean air. The concentrations in the Tedlar bag were well below saturated vapor pressure at room temperature, and complete evaporation was assumed. A gastight syringe was used to inject the sample (
Novel Gas Chromatographic Detector Utilizing the Localized Surface Plasmon Resonance of a Gold Nanoparticle Monolayer inside a Glass Capillary Fong-Yi Chen, Wei-Cheng Chang, Rih-Sheng Jian, and Chia-Jung Lu* Department of Chemistry, National Taiwan Normal University, Taipei, 11677, Taiwan S Supporting Information *
ABSTRACT: This paper presents the design, assembly, and evaluation of a novel gas chromatographic detector intended to measure the absorbance of the localized surface plasmon resonance (LSPR) of a gold nanoparticle monolayer in response to eluted samples from a capillary column. Gold nanoparticles were chemically immobilized on the inner wall of a glass capillary (i.d. 0.8 mm, length = 5−15 cm). The eluted samples flowed through the glass capillary and were adsorbed onto a gold nanoparticle surface, which resulted in changes in the LSPR absorbance. The LSPR probing light source used a green light-emitting diode (LED; λcenter = 520 nm), and the light traveled through the glass wall of the capillary with multiple total reflections. The changes in the light intensity were measured by a photodiode at the rear of the glass capillary. The sensitivity of this detector can be improved by using a longer spiral glass capillary. The detector is more sensitive when operated at a lower temperature and at a slower carrier velocity. The calibration lines of 8 preliminary test compounds were all linear (R2 > 0.99). The detection limits (3σ) ranged from 22 ng (n-butanol) to 174 ng (2-pentanone) depending on the volatility of the chemicals and the affinity to the citrate lignads attached to the gold nanoparticle surface. This detector consumed a very low amount of energy and could be operated with an air carrier gas, which makes this detector a promising option for portable GC or μGC.
I
distance. The relationships between LSPR sensitivity, the distance from the surface, and the refractive index changes can be described using the following equation:20
n the past few decades, the localized surface plasmon resonance (LSPR) of metallic nanoparticles has been extensively studied for possible sensor applications by using its sensitivity toward environmental refractive index changes.1 The LSPR frequencies of many noble metal nanoparticles are within visible to near-IR range, which makes measuring them with a spectrometer simple. One of the most important applications of LSPR sensors is the label-free detection of biological species, which is a great simplification compared with traditional bioassay approaches.2−6 Many researchers are testing the parameters that affect the refractive index sensitivity of LSPR such as particle shape, size, and material.7−13 One particularly important factor is the sensitivity decay with distance from the surface of nanoparticles. Beeram and Zamborini reported a series of studies that used triangular nano-Au to investigate the distancedependent sensitivity and selective binding of biomolecules.14−16 Van Duyne and co-workers have estimated the sensitivity decay on silver nanotriangles using self-assembled monolayer in different chain lengths, and the decay distance for LSPR sensing was determined to be approximately 50 nm.17−19 Kedem et al. used layer-by-layer polymer deposition to determine that the sensing distance of spherical gold nanoparticles ranges between 10 and 20 nm and found the distance to be highly dependent on the nanoparticle size and shape.20 These results have profound implications on developing a LSPR biomolecule sensor because a portion of the large molecules (e.g., proteins, DNA) could be at a low-sensitivity © 2014 American Chemical Society
R = mΔη[1 − exp( −d /l)]
(1)
where R is the sensor response (i.e., wavelength shift or intensity), m is the sensitivity (i.e., Δλ or ΔI per refractive index unit, RIU), Δη is the change in the refractive index (RI), d is the thickness of the dielectric layer, and l is the plasmon decay length. When there is a homogeneous (d = ∞) change in the refractive index surrounding the nanoparticles, such as when they are immersed in a liquid with a given refractive index value, eq 1 can be simplified as R = mΔη. Gas-phase detections using LSPR are inherently difficult for two reasons: first, the adsorption on the surface of nanoparticles is weak for light gases or volatile organic vapors; second, the change in induced refractive indices per adsorbed molecule is small compared to the larger biomolecules (such as proteins) due to the effective volume occupied on the nanoparticle surroundings. In order to facilitate gas/vapor absorption on nanoparticles, Rubinstein and co-workers coated nanoparticles with a polymer film to enhance the vapor detection using LSPR.21 Chen et al. have reported the detection Received: October 1, 2013 Accepted: May 4, 2014 Published: May 4, 2014 5257
dx.doi.org/10.1021/ac4031829 | Anal. Chem. 2014, 86, 5257−5264
Analytical Chemistry
Article
of terpene vapors using thiolate-modified Au nanoparticles.22 Van Duyne’s group demonstrated the use of a high-resolution spectrometer to achieve the detection of gases (i.e., Ar, N2 or He) using the Ag nanotriangles.23 The same group later developed the coating of metal organic framework (MOF) molecules on Ag nanoparticles to enhance the sensitivity to light gases (e.g., CO2, CH4) and found a 14-fold enhancement in the response.24 Using the LSPR of metal oxide-coated nanoparticles for gas sensing has also been reported by Carpenter et al.25,26 In addition, a nanoantenna with a precisely controlled shape and distance has demonstrated superior sensitivity for H2 detection.27 Ibañez and co-workers reported the use of surfactant-stabilized gold nanoparticles as a highly durable LSPR sensor for VOCs.28 Very recently, vapor sensing using interparticle distance modulation in a 3-D plasmonic nanoparticle structure has been demonstrated by Potyrailo et al.29 Fan and Zellers used dual laser reflectance to probe the vapor-induced swelling of a nanoparticle film30 that had evolved from Fan’s previous design of a reflectance-type in-column GC detector.31 Our previous work has focused on studying various types of nanoparticles (e.g., Ag or Au nanospheres or shells) and surface modifications to investigate the selectivity of using LSPR as a VOC sensor array.32−34 In addition, Au nanoparticles have been used as a chemiresistor type of GC detector35,36 and in the stationary phase of a capillary column.37 In this work, we propose a new GC detector design that utilizes the LSPR of Au nanoparticles as a transducer mechanism. Instead of using a regular optical fiber, a reversed configuration (i.e., similar to hollow-fiber optics38,39) was used where gas flow and immobilized Au nanoparticles are inside a glass capillary and the incident light travels along the glass wall of the capillary. In order to minimize the size and improve the portability of the detector, a green light-emitting diode (LED) was used as the light source and a matched range photodiode was used as the light intensity detector. The LSPR detector was evaluated on a benchtop GC under various GC analytical conditions.
GNPs also influenced the gas sensing selectivity in later experiments. Detector Assembly. Both straight and spiral glass capillaries were tested in this study. The spiral glass capillaries were heated and bent by a local glass shop using the same straight glass capillaries obtained from Fishers Scientific. In some cases, the glass capillaries were coated with a silver mirror finish on the outside to maintain light intensity. The coating was performed by sealing the glass capillaries and immersing them in Tollens’ reagent. Amine-stabilized silver ions were gradually reduced by glucose in a basic solution and were used to coat the outer wall of the glass capillary. The photos of both silver-mirror coated and uncoated capillaries are shown in the Supporting Information (Figure S-1). The glass capillary was tightened between two 1/8 in. stainless Swagelok unions using stainless steel nuts and Teflon ferrules (Figure 1a). A short section of deactivated fused silica tube (i.d. 250 μm) was wrapped with a small piece of Teflon tape at one end and inserted into the glass capillary as the sample inlet for the LSPR detector. The deactivated silica tube was bent slightly allowing a green LED to align straight to the cross section of the glass capillary. A plastic syringe was used to
■
EXPERIMENTAL SECTION Nanoparticle Synthesis and Surface Modifications. A standard citric reduction approach has been used to synthesize gold nanoparticles (GNPs) with an average size of approximately 20 nm.32 The inner walls of the glass capillaries (od. = 2.0 mm, id. = 0.8 mm, Fishers Scientific, PA, USA) were first cleaned with Piranha solution and then rinsed with deionized water and dried. (CAUTION: “Piranha” solution reacts violently with organic materials; it must be handled with extreme care.) The glass capillaries were then filled with a solution of 10% (v/v) 3-aminopropyl trimethoxysilane (APTMS) in methanol and sealed for 1 h. After the inner surface of the capillary had been modified with APTMS, the capillary was rinsed with deionized water and filled with an aqueous solution of GNPs. The capillary was stored in a refrigerator for 2 h to allow the GNPs to bind to the glass surface. The preparation and characterization of this surface binding process was previously described by Natan’s group.40 The glass capillary was then rinsed with deionized water and ethanol repeatedly to remove unbounded GNPs and was kept dry for the detector assembly. The citrate that strongly adsorbed on the surface of GNPs as protector during aqueous phase synthesis was not removed completely by this mild rinsing procedure. Therefore, the citrate on the surface of the
Figure 1. (a) Scheme of LSPR-GC detector assembly. Photos of (b) 10 cm straight-type and (c) 15 cm spiral-type detectors. (Note: both pictures show the Ag layer coated on the outside of a glass capillary for better visualization.) 5258
dx.doi.org/10.1021/ac4031829 | Anal. Chem. 2014, 86, 5257−5264
Analytical Chemistry
Article
fix the positions of the LED light and the deactivated fused silica tube. Black vinyl electrical tape was wrapped around the outside of the plastic syringe to block the stray light from the environment. The rear end (i.e., sample exit) of the capillary was kept open and was aligned directly to a photodiode with a built-in filter for the range of the green light (S6429-01, Hamamatsu, Japan). The light was blocked by the fused silica tube at the center of the glass capillary. At the rear of the detector, the photodiode sensed the intensity of a green “halo” of light conducted through the glass wall of the capillary. Figure 1b,c shows the photos of both straight and spiral types of LSPR detectors after assembly had been completed. Instruments and Apparatus. A field emission scanning electron microscope (FESEM, JSM-6500F, JEOL) was used to examine the image of a GNP monolayer on a glass substrate. The SEM image is shown in the Supporting Information (Figure S-2). UV−vis spectra were measured using an Ocean Optics USB-2000 spectrometer. The initial determination of the LSPR spectrum for GNPs and the LSPR spectrum for a vapor response was performed with the same test configuration that was used in our previous report, for which GNPs were bonded to watch-glass slides in the cube cell of a UV−vis spectrometer.32 Test System for the LSPR GC Detector. The LSPR detector was tested on a benchtop gas chromatograph (HP5890, Agilent) equipped with a built-in flame ionization detector. During the comparison experiments with a thermal conductivity detector (TCD), another GC-TCD (Clarus 480, PerkinElmer, MA, USA) was used. A general-purpose DB-5 column (27 m, i.d. 0.53 mm, stationary phase thickness 1 μm) was used throughout all experiments. Test samples were prepared by injecting an aliquot of organic liquid into a Tedlar bag filled with a known volume of clean air. The concentrations in the Tedlar bag were well below saturated vapor pressure at room temperature, and complete evaporation was assumed. A gastight syringe was used to inject the sample (
